Skip to content

tests: derive expected file list from templates/; README: dogfooding note#2

Open
testtest126 wants to merge 1 commit into
mainfrom
claude/code-review-improvements
Open

tests: derive expected file list from templates/; README: dogfooding note#2
testtest126 wants to merge 1 commit into
mainfrom
claude/code-review-improvements

Conversation

@testtest126

Copy link
Copy Markdown
Owner

What

Post-bootstrap code-review pass on the repo. Two focused changes:

  1. tests/install.bats — derive the expected file list from templates/.
    The suite hardcoded a 6-file list (in expected_files() and six explicit diffs), which quietly contradicted the installer's headline property: the copy list is derived from templates/ — drop a file in and it ships automatically, nothing to keep in sync. A new template would have been installed but covered by no test. Replaced both with a template_dest helper + find templates/ derivation, mirroring the installer's adapters/-stripping rule.

  2. README.md — one-line "dogfooded" note, now that the repo runs on its own kit.

No change to install.sh behavior.

Verification

  • shellcheck install.sh → clean
  • bats tests/install.bats → 10/10
  • Proved the refactor's point: in a throwaway copy, a newly-dropped template (adapters/.aider/CONVENTIONS.md) is auto-installed and auto-verified with zero test edits.
  • Proved the gate can go red (principle 8): corrupting the installer so output diverges from templates makes the byte-identical test fail with the exact diff, while the rest stay green.

🤖 Generated with Claude Code

…ding

Code-review pass on the bootstrapped repo:

- tests/install.bats derived the expected file set and the byte-identical
  checks from a hardcoded 6-file list, which silently contradicted the
  installer's headline property (copy list derived from templates/, add a
  file and it ships automatically). Replace both with a `template_dest`
  helper + `find templates/` derivation, so dropping a template in is
  covered end-to-end with nothing to hand-maintain. Verified: a new
  template is auto-covered, and the byte-identical check still goes red on
  a mismatch.
- README: add a one-line "dogfooded" note now that the repo runs its own kit.

No changes to install.sh behavior. shellcheck clean, bats 10/10.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant